Radiology and Nuclear Medicine, Utrecht University and University Medical Center Utrecht, Utrecht, the Netherlands.
Image Sciences Institute, Utrecht University and University Medical Center Utrecht, Utrecht, the Netherlands.
Med Phys. 2020 Jun;47(5):2128-2138. doi: 10.1002/mp.14095. Epub 2020 Mar 10.
It may be challenging to select the optimal scan duration for single-photon emission computed tomography (SPECT) protocols because the activity distribution characteristics can differ in every scan. Using simulations and experiments, we investigated whether the scan duration can be optimized for every scan separately by evaluating the activity distribution during scanning. We refer to this as adaptive scanning.
The feasibility of adaptive scanning was evaluated for the detection of extrahepatic depositions in the pretreatment procedure of radioembolization, in which Tc-labeled macroaggregated albumin ( Tc-MAA) is injected into the liver. We simulated fast 1-min detector rotations and updated the reconstruction with the newly collected counts after every rotation. The scan was terminated when one of the two criteria was met: (a) when the mask difference of the detected extrahepatic deposition between two consecutive rotations was lower than 5%; or (b) when the reconstructed extrahepatic activity was negligible with respect to the total reconstructed activity (<0.075%). The performance of adaptive scanning was evaluated using a digital phantom with various activity distributions, a physical phantom experiment, and simulations based on 129 patient activity distributions.
The digital phantom data showed that the scan termination times substantially depended on the activity distribution characteristics. The experimental phantom data showed the feasibility of adaptive scanning with physical scanner measurements and illustrated that fast detector motion was not limiting the adaptive scanning performance. The patient data showed a large spread in the scan terminations times. By adaptive scanning, the mean scan duration of the patient distributions was shortened from 20 min (current clinical protocol) to 4.8 ± 0.2 min. The detection accuracy of extrahepatic depositions was unaffected and the mean difference in the extrahepatic deposition masks (compared with the 20-min scan) was only 7.0 ± 1.0%.
Our study suggests that the SPECT scan duration can be personalized by assessing the activity distribution characteristics during scanning for the detection of extrahepatic depositions in the pretreatment procedure of radioembolization. The adaptive scanning approach might also be of benefit for other SPECT protocols, as long as a measure of interest is available for optimization.
由于每次扫描的活性分布特征可能不同，因此选择单光子发射计算机断层扫描（SPECT）方案的最佳扫描时间可能具有挑战性。通过模拟和实验，我们通过评估扫描过程中的活性分布来研究是否可以针对每个扫描分别优化扫描时间，我们将此称为自适应扫描。
在放射性栓塞治疗的预处理过程中，评估了自适应扫描检测肝外沉积的可行性，其中将 Tc 标记的聚合白蛋白（Tc-MAA）注入肝脏。我们模拟了快速的 1 分钟探测器旋转，并在每次旋转后用新收集的计数更新重建。当满足以下两个标准之一时，扫描将终止：（a）当两个连续旋转之间检测到的肝外沉积之间的掩模差低于 5％时；或（b）当重建的肝外活性相对于总重建活性可忽略不计（<0.075％）时。使用具有各种活性分布的数字体模，物理体模实验以及基于 129 个患者活性分布的模拟来评估自适应扫描的性能。
数字体模数据表明，扫描终止时间主要取决于活性分布特征。物理体模实验数据表明，自适应扫描具有物理扫描仪测量的可行性，并说明了快速探测器运动不会限制自适应扫描性能。患者数据显示扫描终止时间差异很大。通过自适应扫描，患者分布的平均扫描时间从 20 分钟（当前临床方案）缩短至 4.8±0.2 分钟。肝外沉积的检测准确性不受影响，与 20 分钟扫描相比，肝外沉积掩模的平均差异仅为 7.0±1.0％。
我们的研究表明，通过在放射性栓塞治疗的预处理过程中检测肝外沉积，在扫描过程中评估活性分布特征，可以实现 SPECT 扫描时间的个性化。只要有可供优化的感兴趣的测量值，自适应扫描方法可能对其他 SPECT 方案也有益。
